Kaltura REACH API Guide¶

Kaltura REACH is a governed, budget-controlled marketplace for content enrichment services. It provides a unified API for ordering any enrichment task — from machine-generated captions to professional human translations, AI-powered content moderation to conversational AI agents — against your media entries. Services are delivered by both Machine/AI engines and Human Professional vendors, supporting 80+ languages. Results (caption files, audio tracks, chapters, metadata, moderation reports) are delivered directly back into Kaltura and attached to the entry automatically.

REACH centralizes the management of enrichment services across your organization: credit budgets per account and department, moderation workflows for quality control, content deletion policies for privacy compliance, and vendor abstraction that ensures third-party providers never get direct access to your account. Customers pre-purchase credits that can be used across any integrated service during the contract period.

Base URL: https://www.kaltura.com/api_v3 (may differ by region/deployment)
Auth: KS passed as ks parameter in POST form data (see Session Guide)
Format: Form-encoded POST, format=1 for JSON responses

REACH services can be triggered manually via the API, automatically via REACH profile automation rules (using Boolean event notification conditions), via Kaltura Agents, or through the KMC/KMS management UIs.

Prerequisites¶

Know how to generate Kaltura Sessions (KS) in your backend (see Session Guide)
Have a Kaltura account with REACH services enabled (contact your Kaltura account manager to provision)

Endpoint pattern:

POST https://www.kaltura.com/api_v3/service/{serviceName}/action/{actionName}
Content-Type: application/x-www-form-urlencoded

All parameters are passed as form data. Always include ks=<YOUR_KS> and format=1 (JSON response).

What REACH Provides¶

REACH is an open marketplace — the services listed below are current as of this writing, but the platform is designed to integrate any enrichment service. Each service can be provided by Machine/AI engines, Human Professional vendors, or both, depending on vendor availability for your account.

Captioning & Transcription¶

Machine Captions (ASR) — AI-powered, ~85% accuracy, up to 2-hour turnaround. Supports custom vocabularies/dictionaries for domain-specific terminology.
Professional (Human) Captions — 99%+ accuracy, 3-48 hour turnaround depending on SLA tier. Compliant with closed captioning standards.
Live Captions — Real-time machine-generated captions for scheduled live events.
Alignment — Upload an existing transcript file and REACH syncs it to the audio timeline.

Translation & Localization¶

Translation — Three modes: (1) translate directly from audio, (2) translate from existing captions file, (3) order professional captions first, then translate. 80+ languages supported.
Live Translation — Real-time machine-generated subtitles in a different language during live events.
Dubbing — Machine-based translated audio via text-to-speech, synced to original dialogue timing.

Accessibility¶

Standard Audio Description — Narration of visual elements mixed with original audio.
Extended Audio Description — Pauses video to allow detailed narration of visual content. Uses speech synthesis from time-coded VTT. Requires player plugin.
Sign Language — Creates a child entry with sign language video (ASL or BSL) linked to the parent entry.

Intelligence & Analysis¶

AI Clips — Automatically generate short clips/highlights from longer content using AI analysis. Machine-powered, immediate turnaround.
Quiz — AI-generated quiz questions from video content.
Video Analysis — Visual content analysis including OCR (optical character recognition).
Intelligent Tagging — Automatic topic extraction and semantic tagging from transcript content.
Sentiment Analysis — AI-powered sentiment and tone analysis of spoken content.
Content Moderation — Screen video transcripts and frames against configurable moderation policies using LLMs and computer vision. See Moderation API for policies, rules, and scoring.

Content Enrichment¶

Chaptering — AI-based chapter detection, reviewed by professionals. 24-hour turnaround.
Summary — AI-generated content summary with auto-chaptering.
Metadata Enrichment — Automatic tags, keywords, and metadata extraction.
Document Enrichment — AI-powered enrichment for document-type entries.

Conversational AI¶

Immersive Agent Chat — Text-based conversational AI agent interaction over content.
Immersive Agent Call — Voice-based conversational AI agent interaction over content.

Governance & Billing¶

REACH services are billed in credits, defined per REACH profile. Credit balances, usage, and limits are managed through the REACH profile configuration. Key governance features:

Credit management — Per-account, per-instance (KMS/KAF app), and per-department budgets. Central reporting of usage and billing.
Moderation workflows — Approval/rejection of enrichment task requests with email notifications, controlling when credits are spent.
Content deletion policy — Configure when source content is deleted from vendor systems after processing, ensuring compliance with data privacy regulations (GDPR, DPA).
Vendor abstraction — Third-party vendors receive only time-expired tokens with access to the specific entry being processed, never your account credentials or broader content access.
Automation rules — Define per-channel/category or account-wide rules for automatic enrichment ordering based on content events.

Language Support¶

REACH services support 80+ languages including: English, Spanish, French, German, Portuguese, Arabic, Mandarin Chinese, Cantonese, Japanese, Korean, Russian, Hindi, Hebrew, Italian, Thai, Swedish, Polish, Danish, Finnish, Norwegian, Dutch, Hungarian, Romanian, Vietnamese, Czech, Greek, Indonesian, Turkish, Ukrainian, Welsh, Irish, Catalan, Tamil, Bulgarian, Estonian, Lithuanian, Slovak, Slovenian, Croatian, Farsi, Galician, Malay, Marathi, Mongolian, Taiwanese Mandarin, Afrikaans, Icelandic, Malayalam, Urdu, Zulu, Azerbaijani, Burmese, Bosnian, Georgian, Gujarati, Javanese, Kannada, Kazakh, Macedonian, Nepali, Punjabi, Serbian, Uzbek, Xhosa, Simplified Chinese, Traditional Chinese, and more. Language availability varies by service type and vendor.

Core Concepts¶

Concept	What It Is
REACH Profile	Per-account configuration that defines credit allocation, output preferences (format, moderation, metadata extraction, speaker identification, profanity removal), dictionaries, and content deletion policy.
Vendor Catalog Item	A specific service offering from a vendor — e.g., "English Machine Captions, 2-hour TAT" or "Spanish Human Translation, 5 Business Days". Each has a unique `catalogItemId`.
Entry Vendor Task	A single unit of work: one service applied to one entry. Created by referencing a `catalogItemId` and an `entryId`.

The typical flow: discover available catalog items for your account, then create a task against an entry using the desired catalog item.

Key Enums Reference¶

Service Features (`serviceFeature`)¶

Value	Name	Description
1	CAPTIONS	Speech-to-text captioning (machine ASR or human professional)
2	TRANSLATION	Caption/audio translation
3	ALIGNMENT	Transcript-to-audio alignment
4	AUDIO_DESCRIPTION	Standard audio description narration
5	CHAPTERING	AI chapter detection
6	INTELLIGENT_TAGGING	Automatic topic extraction and semantic tagging
7	DUBBING	Text-to-speech dubbed audio
8	LIVE_CAPTION	Real-time live captions
9	EXTENDED_AUDIO_DESCRIPTION	Extended audio description (pauses video, VTT-based)
10	CLIPS	AI-powered clip generation from longer content
11	LIVE_TRANSLATION	Real-time live translation
12	QUIZ	Quiz generation from video content
13	SUMMARY	Content summarization with auto-chaptering
14	VIDEO_ANALYSIS	Video content analysis (OCR, visual recognition)
15	MODERATION	Content moderation (text + visual policy evaluation)
16	METADATA_ENRICHMENT	Automatic tags, keywords, and metadata extraction
17	SENTIMENT_ANALYSIS	Sentiment and tone analysis of spoken content
18	DOCUMENT_ENRICHMENT	AI enrichment for document-type entries
19	SIGN_LANGUAGE	Sign language video (ASL, BSL)
21	IMMERSIVE_AGENT_CALL	Voice-based conversational AI agent interaction
22	IMMERSIVE_AGENT_CHAT	Text-based conversational AI agent interaction
23	AVATAR_VOD	Pre-recorded avatar video generation from scripts

Service Types (`serviceType`)¶

Value	Name	Description
1	HUMAN	Professional human processing
2	MACHINE	AI/machine processing

Task Statuses (`status`)¶

Value	Name	Description
1	PENDING	Created, waiting to be picked up
2	READY	Ready for vendor processing
3	PROCESSING	Vendor is actively working
4	PENDING_MODERATION	Completed, awaiting review
5	REJECTED	Rejected after moderation
6	ERROR	Processing failed
7	ABORTED	Cancelled
8	PENDING_ENTRY_READY	Waiting for media entry to finish processing
9	SCHEDULED	Scheduled for future execution (live events)

Task Status Flow¶

  Task created ──► PENDING_ENTRY_READY (8)  ──► PENDING (1)
                        (if entry not ready)        │
                                                    │
  Task created ──► PENDING (1) ─────────────────────┤
                                                    │
                        ┌───────────────────────────┤
                        │                           │
                        ▼                           ▼
                   PENDING_MODERATION (4)       READY (2)
                        │                           │
                   ┌────┴────┐                      │
                   ▼         ▼                      ▼
              REJECTED (5)  approved            PROCESSING (3)
                            │                       │
                            ▼                  ┌────┴────┐
                        READY (2) ──►...       ▼         ▼
                                           success    ERROR (6)
                                               │
                                               ▼
                                        PENDING_MODERATION (4)
                                          or READY (2)

  ABORTED (7) ◄── abort() from PENDING or PENDING_MODERATION

Output Formats (`outputFormat`)¶

Value	Name	When to Use
1	SRT	Most widely supported format. Use as the default for web players, video editors, and subtitle tools. Simple text-based timing format.
2	DFXP	XML-based format (TTML subset). Use when you need styling/positioning metadata or for broadcast/TV workflows that require TTML compliance.
3	VTT	WebVTT — the native HTML5 `<track>` format. Use when captions are delivered directly via HTML5 video elements or when you need CSS-based caption styling.

Turn-Around Times (`turnAroundTime`)¶

Value	Name
-1	BEST_EFFORT
0	IMMEDIATE
1-7	1-7 BUSINESS_DAYS
1800	30 minutes
7200	2 hours
10800	3 hours
21600	6 hours
28800	8 hours
43200	12 hours
86400	24 hours
172800	48 hours

Processing Regions (`vendorTaskProcessingRegion`)¶

Value	Name
1	US
2	EU
3	CA

Creation Mode (`creationMode`)¶

Value	Name	Description
1	MANUAL	Created via API or UI by a user
2	AUTOMATIC	Created automatically by a rule or agent

1. List Available Catalog Items¶

Discover which services are available for your account.

vendorCatalogItem.list

POST /api_v3/service/reach_vendorCatalogItem/action/list

ks=<YOUR_KS>
&format=1
&filter[serviceTypeEqual]=2
&filter[serviceFeatureEqual]=1
&filter[statusEqual]=2

Filter parameters

Parameter	Type	Description
`filter[serviceTypeEqual]`	int	Filter by HUMAN (1) or MACHINE (2)
`filter[serviceFeatureEqual]`	int	Filter by feature (1=captions, 2=translation, etc.)
`filter[sourceLanguageEqual]`	string	Filter by source language (e.g., "English"). For best results, also filter client-side to ensure language matches.
`filter[statusEqual]`	int	Filter by status; use 2 (ACTIVE) for available items
`filter[turnAroundTimeEqual]`	int	Filter by TAT
`pager[pageSize]`	int	Results per page (default 30, max 500)
`pager[pageIndex]`	int	Page number (1-based)

Response

{
  "objectType": "KalturaVendorCatalogItemListResponse",
  "objects": [
    {
      "objectType": "KalturaVendorCaptionsCatalogItem",
      "id": 101,
      "vendorPartnerId": 5555,
      "name": "English Machine Captions",
      "serviceType": 2,
      "serviceFeature": 1,
      "sourceLanguage": "English",
      "turnAroundTime": 7200,
      "outputFormat": 1,
      "allowResubmission": true,
      "status": 2
    }
  ],
  "totalCount": 1
}

Save the id — this is the catalogItemId you use when creating tasks.

Catalog item objectType by service feature:

Service Feature	objectType
CAPTIONS (1)	`KalturaVendorCaptionsCatalogItem`
TRANSLATION (2)	`KalturaVendorTranslationCatalogItem`
CHAPTERING (5)	`KalturaVendorChapteringCatalogItem`
DUBBING (7)	`KalturaVendorDubbingCatalogItem`
LIVE_CAPTION (8)	`KalturaVendorLiveCaptionCatalogItem`
CLIPS (10)	`KalturaVendorClipsCatalogItem`
LIVE_TRANSLATION (11)	`KalturaVendorLiveTranslationCatalogItem`
QUIZ (12)	`KalturaVendorQuizCatalogItem`
SUMMARY (13)	`KalturaVendorSummaryCatalogItem`
VIDEO_ANALYSIS (14)	`KalturaVendorVideoAnalysisCatalogItem`
MODERATION (15)	`KalturaVendorModerationCatalogItem`
METADATA_ENRICHMENT (16)	`KalturaVendorMetadataEnrichmentCatalogItem`
DOCUMENT_ENRICHMENT (18)	`KalturaVendorDocumentEnrichmentCatalogItem`

When creating tasks via entryVendorTask.add, always use objectType=KalturaEntryVendorTask regardless of the catalog item type.

2. Get REACH Profile¶

Retrieve your REACH profile to understand account-level settings (credit balance, output preferences, dictionaries).

reachProfile.list

POST /api_v3/service/reach_reachProfile/action/list

ks=<YOUR_KS>
&format=1
&filter[statusEqual]=2

Response

href="#__codelineno-7-1">{ "objects": [ { "id": 1001, "name": "Production Profile", "status": 2, "profileType": 2, "defaultOutputFormat": 1, "enableMachineModeration": 0, "enableHumanModeration": 0, "autoDisplayMachineCaptionsOnPlayer": 1, "autoDisplayHumanCaptionsOnPlayer": 1, "enableMetadataExtraction": 1, "enableSpeakerChangeIndication": 1, "enableProfanityRemoval": 0, "maxCharactersPerCaptionLine": 42, "contentDeletionPolicy": 1, "vendorTaskProcessingRegion": 1, "credit": { "credit": 10000, "fromDate": 1704067200, "toDate": 1735689600 }, "usedCredit": 1250.5, "dictionaries": [], "flavorParamsIds": "" } ], "totalCount": 1 }

Key REACH Profile fields

Field	Description
`defaultOutputFormat`	Default caption format: SRT (1), DFXP (2), VTT (3)
`enableMachineModeration`	If true, machine results go to PENDING_MODERATION before applying
`enableHumanModeration`	If true, human results go to PENDING_MODERATION before applying
`autoDisplayMachineCaptionsOnPlayer`	Auto-display machine captions in the player
`autoDisplayHumanCaptionsOnPlayer`	Auto-display human captions in the player
`enableMetadataExtraction`	Extract tags/keywords from caption content
`enableSpeakerChangeIndication`	Mark speaker changes in captions
`enableProfanityRemoval`	Remove profanity from captions
`maxCharactersPerCaptionLine`	Line length limit for caption formatting
`contentDeletionPolicy`	When to delete source content after processing (1=never, 2=once processed, 3=after week, 4=after month, 5=after 3 months)
`vendorTaskProcessingRegion`	Data residency region: US (1), EU (2), CA (3)
`credit`	Credit allocation and validity period
`usedCredit`	Credits consumed so far
`dictionaries`	Custom vocabularies per language for machine captioning
`flavorParamsIds`	Preferred video flavor IDs for processing (comma-separated)

3. Create a Task (Order a Service)¶

Submit a captioning, translation, or enrichment request for a specific entry.

entryVendorTask.add

POST /api_v3/service/reach_entryVendorTask/action/add

ks=<YOUR_KS>
&format=1
&entryVendorTask[objectType]=KalturaEntryVendorTask
&entryVendorTask[entryId]=1_abc123def
&entryVendorTask[reachProfileId]=1001
&entryVendorTask[catalogItemId]=101

Required parameters

Parameter	Type	Description
`entryVendorTask[entryId]`	string	The media entry ID to process
`entryVendorTask[reachProfileId]`	int	Your REACH profile ID (from step 2)
`entryVendorTask[catalogItemId]`	int	The catalog item ID (from step 1)

Optional parameters

Parameter	Type	Description
`entryVendorTask[notes]`	string	Instructions for human vendors (terminology guidance, context)
`entryVendorTask[context]`	string	Task context identifier (for filtering/grouping)
`entryVendorTask[partnerData]`	string	JSON string with extra data for your own tracking

Response

{
  "id": 98765,
  "partnerId": 12345,
  "entryId": "1_abc123def",
  "status": 1,
  "reachProfileId": 1001,
  "catalogItemId": 101,
  "price": 2.5,
  "createdAt": 1712345678,
  "expectedFinishTime": 1712352878,
  "serviceType": 2,
  "serviceFeature": 1,
  "turnAroundTime": 7200,
  "creationMode": 1,
  "accessKey": "djJ8MTIzNDV8..."
}

The task begins in PENDING (1) status. If the entry is still processing, it starts in PENDING_ENTRY_READY (8) and moves to PENDING once the entry is ready. If moderation is enabled on your REACH profile, completed tasks will go to PENDING_MODERATION (4) before results are applied.

4. Check Task Status¶

entryVendorTask.get

POST /api_v3/service/reach_entryVendorTask/action/get

ks=<YOUR_KS>
&format=1
&id=98765

Response — same structure as the add response, with updated status, finishTime, outputObjectId, and accuracy fields once complete.

Key fields to check:

Field	Description
`status`	Current status (see enum above)
`outputObjectId`	ID of the created output asset (caption ID, flavor ID, cue point ID, etc.)
`accuracy`	Result accuracy percentage (0-100)
`errDescription`	Error details if status is ERROR (6)
`finishTime`	Unix timestamp when processing completed

5. List Tasks¶

Query tasks with filters for monitoring and reporting.

entryVendorTask.list

POST /api_v3/service/reach_entryVendorTask/action/list

ks=<YOUR_KS>
&format=1
&filter[entryIdEqual]=1_abc123def
&filter[statusIn]=1,2,3
&pager[pageSize]=50

Common filter parameters

Parameter	Type	Description
`filter[entryIdEqual]`	string	Filter by entry ID
`filter[statusEqual]`	int	Filter by exact status
`filter[statusIn]`	string	Filter by multiple statuses (comma-separated)
`filter[catalogItemIdEqual]`	int	Filter by catalog item
`filter[reachProfileIdEqual]`	int	Filter by REACH profile
`filter[createdAtGreaterThanOrEqual]`	int	Unix timestamp — tasks created after
`filter[createdAtLessThanOrEqual]`	int	Unix timestamp — tasks created before
`filter[freeText]`	string	Free text search
`pager[pageSize]`	int	Results per page (default 30, max 500)
`pager[pageIndex]`	int	Page number (1-based)

Order by: +createdAt, -createdAt, +updatedAt, -updatedAt, +status, -status, +price, -price, +expectedFinishTime, -expectedFinishTime

6. Manage Tasks¶

Approve a moderated task¶

entryVendorTask.approve

Parameter	Type	Required	Description
`id`	int	Yes	The vendor task ID to approve

POST /api_v3/service/reach_entryVendorTask/action/approve
ks=<YOUR_KS>
&format=1
&id=98765

Moves the task from PENDING_MODERATION (4) to READY (2), applying the results to the entry.

Reject a moderated task¶

entryVendorTask.reject

Parameter	Type	Required	Description
`id`	int	Yes	The vendor task ID to reject
`rejectReason`	string	No	Reason for rejection (logged for audit)

POST /api_v3/service/reach_entryVendorTask/action/reject
ks=<YOUR_KS>
&format=1
&id=98765
&rejectReason=Quality+below+threshold

Moves the task from PENDING_MODERATION (4) to REJECTED (5). Results are discarded and not applied to the entry.

Cancel a pending task¶

entryVendorTask.abort

Parameter	Type	Required	Description
`id`	int	Yes	The vendor task ID to cancel
`abortReason`	string	No	Reason for cancellation

POST /api_v3/service/reach_entryVendorTask/action/abort
ks=<YOUR_KS>
&format=1
&id=98765
&abortReason=No+longer+needed

Works for tasks in PENDING (1) or PENDING_MODERATION (4) status. Aborted tasks are deleted after cancellation; subsequent get calls return ENTRY_VENDOR_TASK_NOT_FOUND.

Response — The aborted task object with status: 7 (ABORTED):

{
  "id": 98765,
  "status": 7,
  "entryId": "1_abc123def",
  "catalogItemId": 101,
  "objectType": "KalturaEntryVendorTask"
}

Export tasks to CSV¶

entryVendorTask.exportToCsv

Parameter	Type	Required	Description
`filter[statusEqual]`	int	No	Filter exported tasks by status (e.g., 2 = READY)
`filter[entryIdEqual]`	string	No	Filter by entry ID
`filter[catalogItemIdEqual]`	int	No	Filter by catalog item
`filter[reachProfileIdEqual]`	int	No	Filter by REACH profile
`filter[createdAtGreaterThanOrEqual]`	int	No	Unix timestamp — tasks created after
`filter[createdAtLessThanOrEqual]`	int	No	Unix timestamp — tasks created before

POST /api_v3/service/reach_entryVendorTask/action/exportToCsv
ks=<YOUR_KS>
&format=1
&filter[statusEqual]=2

Creates a batch job that emails a CSV download link with the filtered tasks.

Full Example: Order Machine Captions for an Entry¶

Step-by-Step curl Example¶

Set up shell variables (see Session Guide for generating a KS):

export KALTURA_SERVICE_URL="https://www.kaltura.com/api_v3"
export KALTURA_PARTNER_ID="YOUR_PARTNER_ID"
export KALTURA_KS="YOUR_KS"

Step 1: Find available machine caption catalog items

curl -X POST "$KALTURA_SERVICE_URL/service/reach_vendorCatalogItem/action/list" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[serviceTypeEqual]=2" \
  -d "filter[serviceFeatureEqual]=1" \
  -d "filter[sourceLanguageEqual]=English" \
  -d "filter[statusEqual]=2"

From the response, note the id of the desired catalog item (e.g., 101). The sourceLanguage filter may not be enforced server-side, so verify the sourceLanguage field in the returned objects matches "English".

Step 2: Get your REACH profile ID

curl -X POST "$KALTURA_SERVICE_URL/service/reach_reachProfile/action/list" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[statusEqual]=2"

From the response, note the id of your active REACH profile (e.g., 1001).

Step 3: Order captions for an entry

curl -X POST "$KALTURA_SERVICE_URL/service/reach_entryVendorTask/action/add" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "entryVendorTask[objectType]=KalturaEntryVendorTask" \
  -d "entryVendorTask[entryId]=1_abc123def" \
  -d "entryVendorTask[reachProfileId]=1001" \
  -d "entryVendorTask[catalogItemId]=101"

From the response, note the task id (e.g., 98765) and expectedFinishTime.

Step 4: Poll for completion

curl -X POST "$KALTURA_SERVICE_URL/service/reach_entryVendorTask/action/get" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=98765"

Poll every 30 seconds until status is 2 (READY) or 6 (ERROR). When READY, the response includes outputObjectId (the caption asset ID) and accuracy (percentage). If ERROR, check errDescription for details.

AI Clips Workflow (Content Lab)¶

AI Clips uses the REACH task system to analyze a video and automatically generate multiple short highlight clips with AI-generated titles, descriptions, and tags. This is exposed in the KMC as "Content Lab".

The workflow has two phases: Generate (AI creates clip suggestions as child entries) and Save (clone a clip into a standalone entry with trimmed content).

Phase 1: Generate AI Clips¶

Order clip generation by creating a REACH task with KalturaClipsVendorTaskData:

entryVendorTask.add

ks=<YOUR_KS>
&format=1
&entryVendorTask[objectType]=KalturaEntryVendorTask
&entryVendorTask[entryId]=1_abc123def
&entryVendorTask[reachProfileId]=1001
&entryVendorTask[catalogItemId]=34832
&entryVendorTask[taskJobData][objectType]=KalturaClipsVendorTaskData
&entryVendorTask[taskJobData][outputLanguage]=English
&entryVendorTask[taskJobData][clipsDuration]=120
&entryVendorTask[taskJobData][instruction]=Generate highlights focusing on key announcements
&entryVendorTask[taskJobData][eventSessionContextId]=

KalturaClipsVendorTaskData fields:

Field	Type	Description
`outputLanguage`	string	Language for generated clip titles/descriptions (e.g., "English")
`clipsDuration`	int	Maximum duration per clip in seconds
`instruction`	string	Free-text instruction for the AI (topic focus, style guidance)
`eventSessionContextId`	string	Optional session context ID for tracking

The task starts in PENDING (1) and moves to PROCESSING (3). Poll with entryVendorTask.get until READY (2). Typical completion time is 30-60 seconds.

Phase 2: List Generated Clips¶

Once the task is READY, the AI has created multiple child entries linked to the source. Retrieve them:

baseEntry.list — find clips by source entry ID:

ks=<YOUR_KS>
&format=1
&filter[rootEntryIdIn]=1_abc123def
&filter[statusIn]=1,2,4
&filter[orderBy]=-createdAt
&pager[pageSize]=500

Each generated clip is a full media entry with: - AI-generated name (clip title) - AI-generated description (clip summary) - AI-generated tags - rootEntryId pointing back to the source entry - Its own duration, thumbnailUrl, etc.

playlist.list — clips may also be grouped in a path playlist:

ks=<YOUR_KS>
&format=1
&filter[referenceIdEqual]=1_abc123def
&filter[playListTypeEqual]=3
&pager[pageSize]=500

Phase 3: Save a Clip as a Standalone Entry¶

To save a specific clip as an independent entry (with its own trimmed video content), use the three-step pattern: clone → trim → set metadata.

Step 1: Clone the source entry

ks=<YOUR_KS>
&format=1
&entryId=1_abc123def
&cloneOptions:item1:objectType=KalturaBaseEntryCloneOptionComponent
&cloneOptions:item1:itemType=5
&cloneOptions:item1:rule=1
&cloneOptions:item2:objectType=KalturaBaseEntryCloneOptionComponent
&cloneOptions:item2:itemType=2
&cloneOptions:item2:rule=1
&cloneOptions:item3:objectType=KalturaBaseEntryCloneOptionComponent
&cloneOptions:item3:itemType=6
&cloneOptions:item3:rule=1
&cloneOptions:item4:objectType=KalturaBaseEntryCloneOptionComponent
&cloneOptions:item4:itemType=7
&cloneOptions:item4:rule=1
&cloneOptions:item5:objectType=KalturaBaseEntryCloneOptionComponent
&cloneOptions:item5:itemType=1
&cloneOptions:item5:rule=1

Clone option itemType values: 1 (content), 2 (access control), 5 (categories), 6 (metadata), 7 (flavors).

Step 2: Trim the cloned entry to clip boundaries

Use media.updateContent with KalturaClipAttributes. Offset and duration are in milliseconds:

ks=<YOUR_KS>
&format=1
&entryId=<CLONED_ENTRY_ID>
&resource[objectType]=KalturaOperationResources
&resource[resources][0][objectType]=KalturaOperationResource
&resource[resources][0][resource][objectType]=KalturaEntryResource
&resource[resources][0][resource][entryId]=1_abc123def
&resource[resources][0][resource][flavorParamsId]=487091
&resource[resources][0][operationAttributes][0][objectType]=KalturaClipAttributes
&resource[resources][0][operationAttributes][0][offset]=350
&resource[resources][0][operationAttributes][0][duration]=205930

To include burned-in captions, add captionAttributes:

&resource[resources][0][operationAttributes][0][captionAttributes][0][objectType]=KalturaRenderCaptionAttributes
&resource[resources][0][operationAttributes][0][captionAttributes][0][captionAssetId]=1_vrrrcg9i

offset — Start position in milliseconds from the source entry
duration — Clip length in milliseconds
flavorParamsId — Get from flavorAsset.list on the source entry (use the original or a ready flavor)
captionAssetId — Optional; get from captionAsset.list to burn captions into the clip

Step 3: Set metadata on the saved clip

ks=<YOUR_KS>
&format=1
&entryId=<CLONED_ENTRY_ID>
&mediaEntry[objectType]=KalturaMediaEntry
&mediaEntry[name]=Introduction to DLI and Its Importance
&mediaEntry[description]=Stephen Wells introduces the session...
&mediaEntry[tags]=DLI, Deep Learning Institute, AI Education
&mediaEntry[displayInSearch]=1

The saved clip will go through server-side transcoding (status 4 = PENDING) and become READY (2) when processing completes.

Step-by-Step curl Example: Full AI Clips Workflow¶

Set up shell variables (see Session Guide for generating a KS):

export KALTURA_SERVICE_URL="https://www.kaltura.com/api_v3"
export KALTURA_PARTNER_ID="YOUR_PARTNER_ID"
export KALTURA_KS="YOUR_KS"

Step 1: Order AI clip generation

curl -X POST "$KALTURA_SERVICE_URL/service/reach_entryVendorTask/action/add" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "entryVendorTask[objectType]=KalturaEntryVendorTask" \
  -d "entryVendorTask[entryId]=1_abc123def" \
  -d "entryVendorTask[reachProfileId]=1001" \
  -d "entryVendorTask[catalogItemId]=34832" \
  -d "entryVendorTask[taskJobData][objectType]=KalturaClipsVendorTaskData" \
  -d "entryVendorTask[taskJobData][outputLanguage]=English" \
  -d "entryVendorTask[taskJobData][clipsDuration]=120" \
  -d "entryVendorTask[taskJobData][instruction]=Generate+highlight+clips"

From the response, note the task id (e.g., 98765).

Step 2: Poll for completion

curl -X POST "$KALTURA_SERVICE_URL/service/reach_entryVendorTask/action/get" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=98765"

Poll every 30 seconds until status is 2 (READY) or 6 (ERROR). Typical completion time for AI clips is 30-60 seconds.

Step 3: List generated clips

curl -X POST "$KALTURA_SERVICE_URL/service/baseEntry/action/list" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[rootEntryIdIn]=1_abc123def" \
  -d "filter[statusIn]=1,2,4" \
  -d "filter[orderBy]=-createdAt" \
  -d "pager[pageSize]=500"

Each object in the response is a generated clip entry with AI-generated name, description, tags, and duration. Note the clip id, duration, and metadata for the next steps.

Step 4: Save a clip as a standalone entry

First, get a source flavor ID for trimming:

curl -X POST "$KALTURA_SERVICE_URL/service/flavorAsset/action/list" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[entryIdEqual]=1_abc123def"

From the response, find a flavor with status = 2 (READY) and note its flavorParamsId (e.g., 487091).

Clone the source entry:

curl -X POST "$KALTURA_SERVICE_URL/service/baseEntry/action/clone" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "entryId=1_abc123def" \
  -d "cloneOptions:item1:objectType=KalturaBaseEntryCloneOptionComponent" \
  -d "cloneOptions:item1:itemType=1" \
  -d "cloneOptions:item1:rule=1"

From the response, note the cloned entry id (e.g., 1_newclone99).

Trim the cloned entry to the clip boundaries (offset and duration in milliseconds):

curl -X POST "$KALTURA_SERVICE_URL/service/media/action/updateContent" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "entryId=1_newclone99" \
  -d "resource[objectType]=KalturaOperationResources" \
  -d "resource[resources][0][objectType]=KalturaOperationResource" \
  -d "resource[resources][0][resource][objectType]=KalturaEntryResource" \
  -d "resource[resources][0][resource][entryId]=1_abc123def" \
  -d "resource[resources][0][resource][flavorParamsId]=487091" \
  -d "resource[resources][0][operationAttributes][0][objectType]=KalturaClipAttributes" \
  -d "resource[resources][0][operationAttributes][0][offset]=0" \
  -d "resource[resources][0][operationAttributes][0][duration]=120000"

Set metadata on the saved clip using the AI-generated values from Step 3:

curl -X POST "$KALTURA_SERVICE_URL/service/media/action/update" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "entryId=1_newclone99" \
  -d "mediaEntry[objectType]=KalturaMediaEntry" \
  -d "mediaEntry[name]=Introduction+to+DLI+and+Its+Importance" \
  -d "mediaEntry[description]=Stephen+Wells+introduces+the+session" \
  -d "mediaEntry[tags]=DLI,+Deep+Learning+Institute,+AI+Education" \
  -d "mediaEntry[displayInSearch]=1"

The saved clip will go through server-side transcoding (entry status 4 = PENDING) and become READY (status 2) when processing completes.

Output Types by Service¶

When a task completes, the result is automatically attached to the entry. The outputObjectId field tells you what was created.

Service Feature	Output Type	How It's Attached
Captions (1)	Caption asset (SRT/DFXP/VTT)	`captionAsset` on the entry, auto-displayed if configured in REACH profile
Translation (2)	Caption asset in target language	`captionAsset` with target language
Alignment (3)	Time-synced caption asset	`captionAsset` aligned to audio
Audio Description (4)	Audio flavor track	`flavorAsset` added to the entry (narration mixed with original audio)
Chaptering (5)	Cue points	`cuePoint` entries marking chapter boundaries
Dubbing (7)	Audio flavor track	`flavorAsset` with dubbed audio in target language
Live Caption (8)	Real-time stream	Captions injected into the live HLS stream via WebSocket
Extended Audio Description (9)	VTT caption asset	`captionAsset` with `usage=1` (EAD), `displayOnPlayer=false`
Clips (10)	Child entries	New media entries linked via `rootEntryId` — each with AI-generated name/description/tags. See AI Clips Workflow
Live Translation (11)	Real-time stream	Translated captions injected into the live stream
Quiz (12)	Cue points	Quiz question cue points attached to the entry
Summary (13)	Metadata	Summary and auto-generated chapters applied to entry metadata
Video Analysis (14)	Metadata	OCR text and visual analysis results applied to entry metadata
Moderation (15)	Metadata	Moderation flags and scores applied to entry metadata
Metadata Enrichment (16)	Tags/keywords	Applied to entry metadata
Document Enrichment (18)	Metadata	Enriched metadata for document entries
Sign Language (19)	Child entry	New media entry linked via `parentEntryId`

REACH Automation Rules¶

REACH profiles support automation rules that automatically order services when specific conditions are met — no application code required. Rules are evaluated by the kReachManager engine whenever content events occur (entry created, entry becomes READY, entry added to category, etc.).

Each rule has conditions (when to trigger) and actions (what to order):

Component	API Object Type	Key Field	Description
Rule	`KalturaRule`	`conditions`, `actions`, `stopProcessing`	Container for one automation rule
Boolean Condition	`KalturaBooleanEventNotificationCondition`	`booleanEventNotificationIds`	References Boolean event notification template IDs that define complex conditions
Category Condition	`KalturaCategoryEntryCondition`	`categoryId`	Triggers when an entry is assigned to a specific category
Action	`KalturaAddEntryVendorTaskAction`	`catalogItemIds`, `entryObjectType`	Auto-orders the specified vendor catalog items

How It Works¶

A content event fires (e.g., entry reaches READY status)
kReachManager loads all active REACH profiles for the partner
For each profile, it evaluates the rules array:
Boolean conditions — retrieves the referenced Boolean event notification templates and evaluates their eventConditions against the event scope
Category conditions — checks if the entry belongs to the specified category
No conditions — rule always triggers
If conditions are met, kReachManager creates EntryVendorTask records for each catalogItemId in the rule's action
Tasks are created with creationMode = 2 (AUTOMATIC)

Configure Rules via Boolean Event Notification Templates¶

Boolean event notification templates (see Webhooks API) define reusable condition logic — "entry tags match a pattern", "entry status changed to READY", "metadata field contains a value". They evaluate conditions and return true/false only — they serve as filters, not dispatchers. REACH rules reference them by ID.

Step 1: Discover available Boolean system templates

curl -X POST "$KALTURA_SERVICE_URL/service/eventnotification_eventnotificationtemplate/action/listTemplates" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[objectType]=KalturaEventNotificationTemplateFilter" \
  -d "filter[typeEqual]=booleanNotification.Boolean"

Step 2: Clone a Boolean template for your account

curl -X POST "$KALTURA_SERVICE_URL/service/eventnotification_eventnotificationtemplate/action/clone" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=$BOOLEAN_SYSTEM_TEMPLATE_ID" \
  -d "eventNotificationTemplate[objectType]=KalturaBooleanNotificationTemplate" \
  -d "eventNotificationTemplate[name]=Entry Ready for Captioning" \
  -d "eventNotificationTemplate[status]=2"

Save the cloned template's id — this is the booleanEventNotificationIds value for the REACH rule.

Step 3: Add the rule to your REACH profile

curl -X POST "$KALTURA_SERVICE_URL/service/reach_reachProfile/action/update" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=$REACH_PROFILE_ID" \
  -d "reachProfile[rules][0][objectType]=KalturaRule" \
  -d "reachProfile[rules][0][description]=Auto-caption when Boolean condition met" \
  -d "reachProfile[rules][0][conditions][0][objectType]=KalturaBooleanEventNotificationCondition" \
  -d "reachProfile[rules][0][conditions][0][booleanEventNotificationIds]=$BOOLEAN_TEMPLATE_ID" \
  -d "reachProfile[rules][0][actions][0][objectType]=KalturaAddEntryVendorTaskAction" \
  -d "reachProfile[rules][0][actions][0][catalogItemIds]=$CATALOG_ITEM_ID" \
  -d "reachProfile[rules][0][actions][0][entryObjectType]=1"

Parameter	Type	Description
`rules[N][objectType]`	string	Always `KalturaRule`
`rules[N][description]`	string	Human-readable rule description
`rules[N][stopProcessing]`	boolean	If true, stop evaluating subsequent rules after this one matches
`rules[N][conditions][M][objectType]`	string	`KalturaBooleanEventNotificationCondition` or `KalturaCategoryEntryCondition`
`rules[N][conditions][M][booleanEventNotificationIds]`	string	Comma-separated Boolean template IDs
`rules[N][actions][M][objectType]`	string	`KalturaAddEntryVendorTaskAction`
`rules[N][actions][M][catalogItemIds]`	string	Comma-separated vendor catalog item IDs to auto-order
`rules[N][actions][M][entryObjectType]`	integer	1=ENTRY (default), 2=ASSET

Category-Based Rules¶

For simpler triggers, use a category condition instead of a Boolean template — auto-order services when an entry is added to a specific category:

curl -X POST "$KALTURA_SERVICE_URL/service/reach_reachProfile/action/update" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=$REACH_PROFILE_ID" \
  -d "reachProfile[rules][0][objectType]=KalturaRule" \
  -d "reachProfile[rules][0][description]=Auto-caption entries in Training Videos category" \
  -d "reachProfile[rules][0][conditions][0][objectType]=KalturaCategoryEntryCondition" \
  -d "reachProfile[rules][0][conditions][0][categoryId]=$CATEGORY_ID" \
  -d "reachProfile[rules][0][actions][0][objectType]=KalturaAddEntryVendorTaskAction" \
  -d "reachProfile[rules][0][actions][0][catalogItemIds]=$CATALOG_ITEM_ID"

Always-On Rules¶

A rule with no conditions triggers for every entry that becomes READY:

curl -X POST "$KALTURA_SERVICE_URL/service/reach_reachProfile/action/update" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "id=$REACH_PROFILE_ID" \
  -d "reachProfile[rules][0][objectType]=KalturaRule" \
  -d "reachProfile[rules][0][description]=Auto-caption all new entries" \
  -d "reachProfile[rules][0][actions][0][objectType]=KalturaAddEntryVendorTaskAction" \
  -d "reachProfile[rules][0][actions][0][catalogItemIds]=$CAPTION_CATALOG_ID,$TRANSLATION_CATALOG_ID"

Multiple catalogItemIds (comma-separated) orders multiple services simultaneously — e.g., captions and translation in one rule.

Multiple Rules with Priority¶

Rules are evaluated in array order. Set stopProcessing=1 to prevent later rules from firing:

  -d "reachProfile[rules][0][objectType]=KalturaRule" \
  -d "reachProfile[rules][0][stopProcessing]=1" \
  -d "reachProfile[rules][0][conditions][0][objectType]=KalturaBooleanEventNotificationCondition" \
  -d "reachProfile[rules][0][conditions][0][booleanEventNotificationIds]=111,222" \
  -d "reachProfile[rules][0][actions][0][objectType]=KalturaAddEntryVendorTaskAction" \
  -d "reachProfile[rules][0][actions][0][catalogItemIds]=555" \
  -d "reachProfile[rules][1][objectType]=KalturaRule" \
  -d "reachProfile[rules][1][conditions][0][objectType]=KalturaCategoryEntryCondition" \
  -d "reachProfile[rules][1][conditions][0][categoryId]=999" \
  -d "reachProfile[rules][1][actions][0][objectType]=KalturaAddEntryVendorTaskAction" \
  -d "reachProfile[rules][1][actions][0][catalogItemIds]=666"

If rule 0's Boolean conditions are met and stopProcessing=1, rule 1 is skipped.

Verifying Automatic Task Creation¶

After triggering an event that matches your rule conditions, verify the task was created automatically:

curl -X POST "$KALTURA_SERVICE_URL/service/reach_entryVendorTask/action/list" \
  -d "ks=$KALTURA_KS" \
  -d "format=1" \
  -d "filter[objectType]=KalturaEntryVendorTaskFilter" \
  -d "filter[entryIdEqual]=$ENTRY_ID" \
  -d "filter[creationModeEqual]=2"

Filter by creationModeEqual=2 (AUTOMATIC) to see only rule-generated tasks. The task's reachProfileId indicates which profile's rule triggered it.

REACH with Kaltura Agents¶

When used through the Agents Manager, REACH tasks are created automatically. The Agents Manager's Action Definitions API returns catalogItemId values that map directly to the REACH vendorCatalogItem IDs described in this guide.

Agents set creationMode = 2 (AUTOMATIC) on tasks they create
The agent handles task creation, status tracking, and retry logic
Results are applied to entries in the same way as manual REACH orders

The Agents Manager supports these REACH services as automated actions: captions, translation, dubbing, summary, moderation, and metadata enrichment. For AI Clips, Quiz, Live Captions, Live Translation, and Video Analysis, use the REACH API directly via entryVendorTask.add. See the Agents Manager guide for the full mapping.

You can mix approaches: use Agents for automated bulk processing, the REACH API directly for on-demand requests, and REACH automation rules for condition-based automatic ordering.

Dictionaries (Custom Vocabularies)¶

For machine captioning, you can configure dictionaries on your REACH profile to improve accuracy for domain-specific terminology.

One dictionary per language
Maximum 8,000 characters
Words/phrases separated by line breaks
Configured via the dictionaries field on the REACH profile
Passed automatically to the machine captioning engine with each task

For human captioning, use the notes field on the task to provide terminology guidance and context to the human editor.

7. Best Practices¶

Use REACH Automation Rules for automatic processing. Set always-on rules or Boolean condition rules on your REACH profile to automatically enrich every new entry — avoids building custom per-entry task submission logic.
Use Agents Manager for multi-step workflows. For "caption → translate → summarize" pipelines, use Agents Manager rather than chaining manual REACH task submissions.
Check credit before bulk operations. Query reachProfile.get to compare credit vs usedCredit before submitting large batches.
Use machine captions (serviceType=2) for speed. Machine captions complete in minutes; human captions take hours. Use machine for time-sensitive content and human for high-accuracy needs.
Poll task status with backoff. Use entryVendorTask.get with increasing intervals. For machine captions (serviceType=2), start polling after 30 seconds with 30-second intervals — tasks typically complete in 1-5 minutes. For human captions (serviceType=1), poll every 15-30 minutes — tasks take hours to days depending on the turn-around time. For AI clips, poll every 30 seconds after submission. Use webhooks on ENTRY_VENDOR_TASK events instead of polling when possible.
Use creationModeEqual=2 filter to distinguish auto-created tasks (from automation rules) from manually submitted tasks in reports.
Use webhooks for task completion notifications. Set up an HTTP webhook on ENTRY_VENDOR_TASK events rather than polling for task status changes.
Configure dictionaries for domain-specific content. Add custom vocabulary (product names, technical terms) to improve machine captioning accuracy.

8. Error Handling¶

Scenario	What Happens
Entry not ready yet	Task starts in PENDING_ENTRY_READY (8), auto-transitions to PENDING (1) when ready
Vendor processing fails	Task status becomes ERROR (6), check `errDescription` for details
Credit limit exceeded	Task creation fails with an API error
Access key expired	Vendor-side issue; key auto-extends based on catalog item TAT
KS expired / invalid	API returns 401/403; generate a new KS and retry
Kaltura API error	Response `objectType` is `KalturaAPIException` with `code` and `message` fields

Common API Error Codes¶

Code	Meaning
`ENTRY_VENDOR_TASK_NOT_FOUND`	Task was cancelled/aborted (and deleted) or never existed
`ENTRY_NOT_READY`	Cannot order a task for an entry that has no content or is still processing
`SERVICE_ACCESS_CONTROL_RESTRICTED`	Missing or incorrect `serviceUrl` configuration
`INVALID_KS`	KS is expired, malformed, or lacks required privileges

Retry strategy: For transient errors (HTTP 5xx, timeouts), retry with exponential backoff: 1s, 2s, 4s, with jitter, up to 3 retries. For client errors (INVALID_KS, ENTRY_VENDOR_TASK_NOT_FOUND, credit limit exceeded), fix the request before retrying — these will not resolve on their own. For async operations (task processing), poll with increasing intervals (5s, 10s, 30s) rather than tight loops.

Session Guide — Generate the KS required for all REACH API calls
AppTokens — Secure server-to-server auth for production REACH integrations
Upload & Delivery — Upload content before ordering REACH enrichment services
eSearch — Search entries enriched by REACH-generated captions, metadata, and tags
AI Genie — Conversational AI that benefits from REACH-generated transcripts and enrichment
Player Embed — Play entries with REACH-generated captions, translations, and audio descriptions
Agents Manager — Automate REACH tasks with event-driven agents
Webhooks API — Boolean templates as REACH automation rule conditions; HTTP callbacks for task completion events
Captions & Transcripts API — Caption assets created and managed by REACH tasks; format conversion, multi-language workflows
Custom Metadata API — Custom metadata schemas for content enrichment
Analytics Reports — Monitor REACH task volumes and costs via reports
Moderation API — AI moderation via REACH (serviceFeature=15), policies, rules, and category auto-action
Content Lab — Unisphere widget UI for REACH enrichment (summaries, chapters, clips, quizzes)
Agents Widget — Management UI for automated REACH enrichment tasks
Events Platform — Auto-enrich virtual event recordings with REACH services
Captions Editor — Interactive editing of REACH-generated captions
Multi-Stream — Apply REACH enrichment to parent and child stream entries
VOD Avatar — Enrich generated avatar videos with REACH services
Cue Points & Interactive Video — Chapters, quizzes, and clips generated by REACH services are stored as cue points
Quiz API — Quiz questions generated by REACH / Content Lab
Chapters & Slides API — Chapters and slides generated by REACH enrichment

Kaltura REACH API Guide¶

Prerequisites¶

What REACH Provides¶

Captioning & Transcription¶

Translation & Localization¶

Accessibility¶

Intelligence & Analysis¶

Content Enrichment¶

Conversational AI¶

Governance & Billing¶

Language Support¶

Core Concepts¶

Key Enums Reference¶

Service Features (serviceFeature)¶

Service Types (serviceType)¶

Task Statuses (status)¶

Task Status Flow¶

Output Formats (outputFormat)¶

Turn-Around Times (turnAroundTime)¶

Processing Regions (vendorTaskProcessingRegion)¶

Creation Mode (creationMode)¶

1. List Available Catalog Items¶

2. Get REACH Profile¶

3. Create a Task (Order a Service)¶

4. Check Task Status¶

5. List Tasks¶

6. Manage Tasks¶

Approve a moderated task¶

Reject a moderated task¶

Cancel a pending task¶

Export tasks to CSV¶

Full Example: Order Machine Captions for an Entry¶

Step-by-Step curl Example¶

AI Clips Workflow (Content Lab)¶

Phase 1: Generate AI Clips¶

Phase 2: List Generated Clips¶

Phase 3: Save a Clip as a Standalone Entry¶

Step-by-Step curl Example: Full AI Clips Workflow¶

Output Types by Service¶

REACH Automation Rules¶

How It Works¶

Configure Rules via Boolean Event Notification Templates¶

Category-Based Rules¶

Always-On Rules¶

Multiple Rules with Priority¶

Verifying Automatic Task Creation¶

REACH with Kaltura Agents¶

Dictionaries (Custom Vocabularies)¶

7. Best Practices¶

8. Error Handling¶

Common API Error Codes¶

9. Related Guides¶

Service Features (`serviceFeature`)¶

Service Types (`serviceType`)¶

Task Statuses (`status`)¶

Output Formats (`outputFormat`)¶

Turn-Around Times (`turnAroundTime`)¶

Processing Regions (`vendorTaskProcessingRegion`)¶

Creation Mode (`creationMode`)¶